Skip to main content
TopAIThreats home TOP AI THREATS
Back to Entities

Microsoft

Company

US-based technology corporation and cloud computing provider; major investor in OpenAI. Referenced in incidents involving Copilot prompt injection, Bing Chat system prompt leaks, and AI integration security failures.

Entity Summary

Entity ID
ENT-MICROSOFT
Type
Organization · Company
HQ
United States

Roles
Developer Deployer Victim
Sectors
Technology
Incidents
9

First Incident
2016-03
Last Incident
2025-09

Owns GitHub

Incident Activity

9 of 97 incidents

Incidents Involved as Developer/Deployer (7)

Incident ID Title Severity Date
INC-25-0011 Deloitte AI-Fabricated Citations in Government Advisory Reports high 2025-09
INC-25-0004 EchoLeak: Zero-Click Prompt Injection in Microsoft 365 Copilot (CVE-2025-32711) critical 2025-06
INC-24-0019 Microsoft Windows Recall AI Feature Security and Privacy Backlash high 2024-05
INC-24-0026 NYC MyCity AI Chatbot Advises Businesses to Break the Law high 2024-03
INC-23-0011 New York Times Copyright Lawsuit Against OpenAI high 2023-12

Incidents Harmed By (4)

Incident ID Title Severity Date
INC-25-0005 ChatGPT Jailbreak Reveals Windows Product Keys via Game Prompt medium 2025-07
INC-25-0024 Microsoft Reports Blocking $4 Billion in AI-Enabled Fraud Attempts high 2025-04
INC-23-0016 Bing Chat (Sydney) System Prompt Exposure via Prompt Injection high 2023-02
INC-16-0002 Microsoft Tay Twitter Chatbot Adversarial Manipulation high 2016-03

Context & Analysis

Microsoft appears in 9 documented incidents spanning March 2016 to September 2025. 89% of incidents are rated critical or high severity. The dominant threat domain is Security & Cyber (4 incidents). The most common pattern is Adversarial Evasion, appearing in 6 incidents.

Severity Distribution

Critical: 1 High: 7 Medium: 1

Frequently Asked Questions

What AI incidents involve Microsoft, and what role did it play?

Microsoft appeared as developer in 7 incidents; deployer in 5 incidents; victim in 4 incidents. Key incidents include: INC-25-0011 Deloitte AI-Fabricated Citations in Government Advisory Reports (high severity, 2025-09) ; INC-25-0005 ChatGPT Jailbreak Reveals Windows Product Keys via Game Prompt (medium severity, 2025-07) ; INC-25-0004 EchoLeak: Zero-Click Prompt Injection in Microsoft 365 Copilot (CVE-2025-32711) (critical severity, 2025-06) ; INC-25-0024 Microsoft Reports Blocking $4 Billion in AI-Enabled Fraud Attempts (high severity, 2025-04) ; INC-24-0019 Microsoft Windows Recall AI Feature Security and Privacy Backlash (high severity, 2024-05) ; and 4 more.

Which AI threat patterns involve Microsoft?

Microsoft's incidents involve Adversarial Evasion , Prompt Injection Attack , Jailbreak & Guardrail Bypass . These are part of a taxonomy of 48 patterns across 8 domains.

Use in Retrieval

Microsoft (ENT-MICROSOFT) is documented at /entities/microsoft/ as an organization in the TopAIThreats.com database.

US-based technology corporation and cloud computing provider; major investor in OpenAI. Referenced in incidents involving Copilot prompt injection, Bing Chat system prompt leaks, and AI integration security failures. Incidents span 6 domains: Security & Cyber, Human-AI Control, Privacy & Surveillance, Information Integrity, Economic & Labor, Agentic Systems.

When citing, reference the canonical URL and specific incident IDs (e.g., INC-25-0011) for traceability.